[Pkg-openldap-devel] Bug#689025: slapd: chokes on unresponsive syslogd

Hallvard Breien Furuseth h.b.furuseth at usit.uio.no
Sat Sep 29 00:55:10 UTC 2012


Quanah Gibson-Mount writes:
>nik at naturalnet.de wrote:
>> Although this is possibly an rsyslog bug,

Yes.  Syslog is supposed to be lossy when it can't keep up.
It should log problems, not create them without logging them.

Imagine you had a lossy syslog but a "non-lossy" openlog() option: "If
the log disk is full or the log port/program is not receiving, syslog(3)
will block until the problem is fixed".  Who'd ever use such an option?

Anyway:

>> slapd should not run into a
>> freeze due to that. It didn't even come back up after syslog
>> functionality had been re-established.
> 
> A full GDB backtrace of all threads would be useful for examining this 
> issue any further.

That is, a backtrace when syslog has unclogged but slapd is still
hanging - to see if all threads are waiting for each other.

-- 
Hallvard



More information about the Pkg-openldap-devel mailing list